Insiu de Recerca en Economia Aplicada Regional i Pública Documen de Treball 2014/24 1/26 Research Insiue of Applied Economics Working Paper 2014/24 1/26 European governmen bond marke inegraion in urbulen imes Pilar Abad and Helena Chuliá
4 WEBSITE: www.ub.edu/irea/ CONTACT: irea@ub.edu The Research Insiue of Applied Economics (IREA) in Barcelona was founded in 2005, as a research insiue in applied economics. Three consolidaed research groups make up he insiue: AQR, RISK and GiM, and a large number of members are involved in he Insiue. IREA focuses on four prioriy lines of invesigaion: (i) he quaniaive sudy of regional and urban economic aciviy and analysis of regional and local economic policies, (ii) sudy of public economic aciviy in markes, paricularly in he fields of empirical evaluaion of privaizaion, he regulaion and compeiion in he markes of public services using sae of indusrial economy, (iii) risk analysis in finance and insurance, and (iv) he developmen of micro and macro economerics applied for he analysis of economic aciviy, paricularly for quaniaive evaluaion of public policies. IREA Working Papers ofen represen preliminary work and are circulaed o encourage discussion. Ciaion of such a paper should accoun for is provisional characer. For ha reason, IREA Working Papers may no be reproduced or disribued wihou he wrien consen of he auhor. A revised version may be available direcly from he auhor. Any opinions expressed here are hose of he auhor(s) and no hose of IREA. Research published in his series may include views on policy, bu he insiue iself akes no insiuional policy posiions.
Absrac In his paper we invesigae he dynamics of European governmen bond marke inegraion during he financial crisis and, subsequenly, during he European sovereign deb crisis. Based on he approach developed by Bae e al. (2003), we adop an inuiive measure of inegraion: he higher he number of join exreme price rises or falls (coexceedances), he higher he degree of inegraion. We also analyse he underlying deerminans of he dynamics of inegraion using a binomial logisic regression. Our resuls reveal ha he level of inegraion of European governmen bond markes wih he euro area has changed over ime, wih noable differences beween he financial and he European sovereign deb crises. We find ha he Euribor, unexpeced moneary policy announcemens from he ECB and boh regional and inernaional volailiy play an imporan role in deermining he level of inegraion, and ha, in general, he relevance of hese facors does no change beween he financial and he sovereign de crises. JEL classificaion: C25; F36; G15 Keywords: Financial inegraion; European governmen bond markes; coexceedances; exreme reurns; logisic regression. Pilar Abad: Deparmen of Economic, Universidad Rey Juan Carlos & Riskcener-IREA(Barcelona, Spain) (pilar.abad@urjc.es). Helena Chuliá: Deparmen of Economerics & Riskcener-IREA, Universia de Barcelona (Barcelona, Spain) (hchulia@ub.edu). Acknowledgemens This work has been funded by he Spanish Minisry of Economy and Compeiiveness (ECO2011-23959 and ECO2012-35584).
1. Inroducion The exen o which European governmen bond markes are inegraed is a key quesion for policymakers and marke paricipans. Policymakers are paricularly keen o undersand he mechanisms ha link hese markes in order o be able o make effecive moneary policy decisions and o mainain financial sabiliy. Likewise, an undersanding of bond marke linkages can help marke paricipans formulae appropriae risk managemen sraegies and invesmen decisions. This ineres becomes even greaer in years of urmoil when financial markes are hi by exreme shocks. In he financial lieraure, a wide variey of frameworks have been employed for he empirical examinaion of he inegraion of European governmen bond markes. The cenral focus of early papers was on he role ha he European Moneary Union (EMU) played in he process of financial inegraion of he EU-15 bond markes. In his line of research, some sudies have assessed he relaive imporance of sysemic and idiosyncraic risk in EMU sovereign yield spreads. Geyer e al. (2004) and Pagano and von Thadden (2004) find ha yield differenials under EMU are driven mainly by a common risk (defaul) facor and sugges ha liquidiy differences play, a bes, only a minor role in he ime series behaviour of yield spreads. Gomez-Puig (2009a and 2009b) presens evidence o he effec ha i was domesic, raher han inernaional, risk facors ha were he primary drivers of en-year yield spread differenials over Germany in all EMU counries in he seven years following he iniiaion of moneary inegraion. A differen perspecive is provided by Chrisiansen (2007) who conducs a volailiy-spillover analysis o show ha he bond markes of EMU counries are more inegraed han hose of non-emu counries and ha hese markes became more inegraed following he inroducion of he euro. A more recen sudy by Beber e al. (2009) finds ha he bulk of sovereign yield spreads can be explained by differences in credi qualiy, hough liquidiy plays a non-rivial role. Finally, Abad e al. (2010) find ha euro area counries are only parially inegraed and presen differences in heir marke liquidiy and defaul risk. They also find ha he markes of counries sharing a moneary policy are more vulnerable o regional risk facors and ha he counries ha oped o say ou of he Moneary Union are more vulnerable o global risk facors.
Anoher line of empirical research of European governmen bond marke inegraion incorporaes he new EU members ino he analysis. 1 Drawing on a se of complemenary echniques, including dynamic coinegraion and ime-varying correlaions, Kim e al. (2005) find ha he degree of inegraion of he new members wih he German bond marke is weak and sable, wih lile evidence of any furher srenghening despie increased poliical inegraion. Wihin he framework of a facor model for marke reurns, Cappiello e al. (2006) only documen an increase in inegraion for he Czech Republic s bond marke versus Germany s. Finally, a new line of research invesigaes he impac of he financial crisis on European governmen bond marke inegraion. Von Hagen e al. (2011) find ha he larger spreads observed during he financial crisis are he resul of a higher penaly imposed by he markes on fiscal imbalances and of greaer inernaional risk aversion, i.e. a higher common risk facor in he spreads. Pozzi and Wolswijk (2012) and Abad e al. (2014) exploi he implicaions of asse pricing models o analyse he effecs of he financial crisis. The resuls of Pozzi and Wolswijk (2012) sugges ha he idiosyncraic facors were almos eliminaed in all counries by 2006 bu subsequenly reappeared as a consequence of he financial crisis. Abad e al. (2014) show ha, from he onse of he financial marke ensions in Augus 2007, markes moved owards higher segmenaion, and he differeniaion of counry risk facors increased subsanially across counries. Chrisiansen (2014), who measures he inegraion of European governmen bond markes employing he explanaory power of facor models, concludes ha he inegraion of EMU members has no been so grea during he recen crisis. This sudy assesses he inegraion of a seleced number of European governmen bond markes wih he euro area during periods of urbulence, when invesors and policymakers have a paricularly srong ineres in wheher and how shocks propagae o oher counries. Following he launch of he euro in January 1999, he markes priced he deb of he European member saes as being virually idenical. In he period 2003-2007, spreads remained very small and did no reflec differences in he fiscal posiions of he counries. 2 As such, he period was characerised by a significan underpricing of risk, leaving invesors o search for yield in an environmen of abundan global liquidiy. This progress owards 1 The new members included are usually he Czech Republic, Hungary and Poland given ha hey are he only ones wih sufficienly developed bond markes. 2 Cassola and Morana (2012) also poin ou ha a peculiar feaure of he pre-crisis euro area money marke was he virual absence of EURIBOR-Overnigh Index Swaps spreads.
financial inegraion was inerruped and reversed, however, by he global financial crisis and, more recenly, by he European sovereign deb crisis, in which sovereign bond markes have been dominaed by sharp differeniaion, especially across borders. Based on he approach developed by Bae e al. (2003), we measure he inegraion of each European governmen s bond marke wih he euro area by examining how ofen exreme reurns on each bond marke and he euro area occur simulaneously. This analysis provides helpful informaion on he dynamics leading o join exreme price rises or falls and allows us o adop an inuiive measure of inegraion: he higher he number of coexceedances wih he euro area, he higher he degree of inegraion. Bae e al. (2003) capure he coincidence of exreme reurn shocks across counries wihin a broader region and also across regions. They define conagion wihin regions as he fracion of coexceedances ha canno be explained by fundamenals and conagion across regions as he fracion of coexceedances unexplained by fundamenals ha is explained by he exceedances from oher counries. 3 This approach is used by Chrisiansen and Ranaldo (2009) o analyse he financial inegraion of he sock markes in he en new EU member saes from he former Communis counries of Easern and Cenral Europe as well as he inegraion of wo groups of counries, namely, new and old member saes. In his paper, we are ineresed in analysing he inegraion of a seleced number of European governmen bond markes wih he euro area and in esing wheher here are differences across counries wih respec o he underlying deerminans of he level of inegraion. To his end, in a firs sep, we carry ou a hierarchical cluser analysis ha allows us o group counries in erms of heir level of inegraion over he sample period. In a second sep, we use a logisic regression model o deermine he underlying deerminans of he observed dynamics of inegraion. We address wo basic ses of quesions. Firs, how closely are he European governmen bond markes associaed wih he euro area? And, has he level of inegraion of hese markes changed during he recen years of urmoil? I is inuiive ha financial marke inegraion changes wih economic condiions. Second, which facors are associaed wih an increase (decrease) in he probabiliy of observing exreme reurns across markes? Have he effecs of hese facors changed during he financial crisis and, subsequenly, during he 3 Their approach possesses wo advanages. Firs, conrary o sandard correlaion measures, i is robus o ime-varying volailiy and deparure from normaliy. Second, he correlaion coefficien is a linear measure, which is inappropriae for analysing nonlinear phenomena, as financial marke inegraion poenially migh be (see Baur and Schulze, 2005; Dungey and Marin, 2007).
European sovereign deb crisis? And is he level of inegraion of European governmen bond markes driven by global (US) or regional (EMU) facors? The main resuls of his paper can be summarized as follows. Firs, we find ha he level of inegraion of European governmen bond markes wih he euro area has changed over ime and ha he bond markes analysed group differenly over he sample period in erms of inegraion. Second, our analysis of he facors affecing he inegraion of European governmen bond markes shows ha: (i) here is a subsiuion effec beween he bond marke and money marke insrumens ha leads o a decrease in he level of inegraion, (ii) inegraion increases in highly volaile periods in boh regional and inernaional sock markes, (iii) unexpeced news releases from he ECB increase uncerainy and decrease he level of inegraion of European governmen bond markes, (iv) in general, he relevance of hese facors does no change during he financial and he sovereign deb crises and, (v) he new members are hose ha behave mos differenly in erms of he facors associaed wih he level of inegraion. The res of his paper is organized as follows. In Secion 2 we presen our daa. In Secion 3 we invesigae he evoluion of he inegraion of European governmen bond markes wih he euro area. In Secion 4 we examine he deermining facors of European governmen bond marke inegraion. Finally, we conclude in Secion 5. 2. Daa The daa consis of he en-year JPMorgan Governmen Global Bond Index (JPMGBI), expressed in erms of a common currency, he euro, and he sample includes 16 European counries. Our sudy focuses on en EMU EU-15 counries (Ausria, Belgium, France, Germany, Greece, Ireland, Ialy, he Neherlands, Porugal and Spain) 4 and six non-emu counries (Denmark, he Czech Republic, Hungary, Poland, Sweden and he UK). As a proxy for he enire euro area we use he JP Morgan EMU Governmen Index. These bond marke indices are ransformed ino reurns by aking he firs difference of he naural log of each bond price index. All daa have been colleced from Thomson Daasream. We use daily daa for he period January 2005 hrough December 2013, hus our sample covers he recen years of urmoil (iniially he financial crisis and, subsequenly, he 4 Finland is no included in he sudy due o a lack of available daa.
European sovereign deb crisis). We define he saring poin of he financial crisis as Augus 2007, when equiy markes iniially fell and cenral banks sared inervening o provide liquidiy o financial markes. For our analysis, we mach he end of he financial crisis wih he beginning of he European sovereign deb crisis. As poined ou by Chrisiansen (2014), daing he European sovereign deb crisis is no a sraighforward ask as no official daes are available. Generally, i is considered o have begun in lae 2009 and is sill running is course. Therefore, we define he saring poin of he sovereign crisis as January 2010 and i runs ill he end of our daase in December 2013. Following Bae e al. (2003), we define an exreme reurn, or exceedance, as one ha lies eiher below (above) he 5h (95h) quanile of he marginal reurn disribuion. Similarly, we define a coexceedance as he occurrence of exreme reurns in one European governmen bond marke and in he euro area simulaneously on a given day; hus, he higher he number of coexceedances wih he euro area, he higher he degree of inegraion. We rea posiive exreme reurns separaely from negaive exreme reurns as some auhors sugges an asymmeric effec of explanaory variables on he ails of he reurn disribuion (see Bae e al., 2003 and Crisiansen and Ranaldo, 2009). Therefore, for each European governmen bond marke we disinguish beween hree evens: negaive coexceedance wih he euro area for a given day, posiive coexceedance wih he euro area for a given day and no coexceedance wih he euro area for a given day. Table 1 shows he relaive frequency of he join occurrences of exreme reurns beween each European governmen bond marke and he euro area on a paricular day. We compue he number of coexceedances for he enire sample, for he ranquil period (from 1 January 2005 o 6 Augus 2007), he financial crisis (from 7 Augus 2007 o 31 December 2009) and, he European sovereign crisis (from 1 January 2010 o 15 December 2013). As is sandard in he lieraure, we have divided he European counries ino four groups: (1) EMU EU-15 cenral counries, (2) EMU EU-15 peripheral counries, (3) Non- EMU new EU counries, and (4) non-emu EU-15 counries. In our sudy, hese groups are composed of he following counries: EMU EU-15 cenral (Ausria, Belgium, France, Germany and he Neherlands), EMU EU-15 peripheral (Greece, Ireland, Ialy, Porugal and Spain), Non-EMU new EU (he Czech Republic, Hungary and Poland), and non-emu EU-15 (Denmark, Sweden and he UK). The number of coexceedances of cenral and peripheral bond markes wih he euro area is lower during he ranquil period han during he crisis periods and, wihin he crisis periods, i is higher during he financial crisis. The
ANOVA es, repored in Table 2, in general confirms he saisical significance of hese differences in he number of coexceedances suggesing ha he level of inegraion of hese bond markes wih he euro area, measured by he number of coexceedances, differs across he hree sub-samples. 5 The disribuion of negaive and posiive coexceedances is largely symmerical. In general, here are no grea differences in he number of coexceedances across he cenral bond markes; however, in he case of peripheral bond markes, Ialy is he mos frequen paricipan in coexceedance evens in he hree sub-samples, ogeher wih Spain during he periods of crisis, indicaing ha hese are he mos highly inegraed peripheral bond markes wih he euro area. Likewise, Greece is he counry wih he lowes number of coexceedances wih he euro area during he sovereign deb crisis, suggesing ha his is he leas inegraed of he peripheral bond markes during his period. The picure is somewha differen when we look a he new members. When we disinguish beween posiive and negaive coexceedances, he ANOVA es does no rejec he null hypohesis ha he number of coexceedances is equal in he hree sub-samples and in he wo crisis periods considered, indicaing ha heir level of inegraion wih he euro area has no changed over he sample period. As wih he cenral and peripheral bond markes, he disribuion of coexceedances is mosly symmerical bu, in general, he number of coexceedances of he new members bond markes and he euro area is lower, suggesing ha he new members are less inegraed wih he euro area han are he cenral and peripheral bond markes. 6 Finally, in erms of he number of coexceedances over he sample period, Denmark, Sweden and he UK presen a similar behaviour o ha of he cenral and peripheral bond markes. The ANOVA es indicaes ha he number of coexceedances of Denmark and Sweden wih he euro area is significanly higher during he financial crisis han during he sovereign deb crisis. 2.1. Explanaory variables We examine four main hypoheses relaing marke condiions o he likelihood of coexceedances and, o his end, we use a large se of explanaory variables. Firs, several argumens such as he fligh-o-qualiy proposed by Caballero and Krishnamurhy (2008) and he liquidiy spirals proposed by Brunnermeier and Pedersen (2009) sugges a subsiuion effec beween equiies, money marke insrumens and bonds in urbulen 5 To es he equaliy in he number of coexceedances across he ranquil, financial crisis and European sovereign deb crisis periods, we carry ou he ANOVA es wih he null hypohesis ha he number of coexceedances is equal in he hree sub-samples and in he wo crisis periods considered. 6 Chrisiansen and Ranaldo (2009) come o he same conclusion for European sock markes.
periods. Capial flows owards oher markes migh weaken he inegraion of European governmen bond markes. To es his hypohesis we include daily reurns of he European sock marke (Eurosoxx50), he relevan index for each local marke 7 and he hree-monh inerbank ineres rae (Euribor). 8 Second, according o Crisiansen and Ranaldo (2009) inegraion or he propagaion of shocks is more likely in a highly volaile environmen. Thus, he hypohesis o be esed is wheher inegraion is srenghened when volailiy is pervasively high in he financial markes. As a proxy of European financial marke volailiy we use he European sock marke volailiy. 9 As is sandard in he lieraure, we compue volailiy as he square roo of he condiional variance esimaed using an AR(1)-GARCH(1,1) model. Third, as poined ou by Manganelli and Wolswijk (2009), he greaer inegraion of European governmen bond markes mainly reflecs he progressive eliminaion of uncerainy in he euro area. Similarly, as suggesed by Abad and Chuliá (2013), unexpeced moneary policy announcemens from he European Cenral Bank (ECB) increase uncerainy and decrease he level of inegraion of European governmen bond markes wih he euro area. Thus, if moneary policy announcemens surprise he markes and generae uncerainy, his could weaken inegraion. To es his hypohesis, we include he surprise or he unexpeced componen of he news announcemens 10 released by he ECB. Finally, wih he aim of disinguishing regional facors from global facors, he fourh group of variables is associaed wih he US. These variables are he reurn of he US sock marke (S&P500 Composie index), he hree-monh Treasury bill rae, he US sock reurn volailiy 11 and he surprise or he unexpeced componen of he news announcemens released by he Federal Reserve (Fed). The hypoheses o be esed are (i) wheher here is a 7 The relevan indexes are he ATX index for Ausria, he BEL 20 index for Belgium, he CAC 40 index for France, he PX index for he Czech Republic, he OMXC20 index for Denmark, he DAX 30 index for Germany, he ATHEX Composie index for Greece, he BUX index for Hungary, he ISEQ index for Ireland, he FTSE MIB index for Ialy, he AEX index for he Neherlands, he WSE index for Poland, he PSI-20 for Porugal, he IBEX 35 for Spain, he OMXS30 index for Sweden, and he FTSE 100 for he UK. 8 The Euribor is included in firs differences because a uni roo canno be rejeced. 9 To avoid he problem of so-called complee separaion when esimaing he binomial logi regression, we do no include he volailiy of European governmen bond markes as an explanaory variable. 10 An imporan common finding in he exan lieraure is ha only he surprise componen of moneary policy has a significan effec on asse reurns, whereas he effec of expeced policy acions is saisically insignifican (see Bomfim, 2003, and Bernanke and Kuner, 2005, among ohers). 11 The correlaion beween he US and he European sock reurn volailiies during our sample period is 0.9. Wih he aim of avoiding he mulicollineariy problem, we proceed as follows. Firs, we calculae he US sock reurn volailiy as he square roo of he condiional variance esimaed using an AR(1)-GARCH(1,1) model. Then, we remove he influence of he European sock reurn volailiy by running a regression of he US sock reurn volailiy on European sock reurn volailiy. Finally we ake he residuals of his regression as our proxy for he US sock reurn volailiy.
subsiuion effec beween European and US asses similar o ha beween European asse classes, (ii) wheher higher volailiy in inernaional financial markes increases he level of inegraion and, (iii) wheher increasing uncerainy in he US (measured hough moneary policy surprises) leads o a decrease in he level of inegraion, as suggesed by Abad and Chuliá (2013). To obain a measure of he surprise in he Fed announcemens we use he mehodology proposed by Kuner (2001). For an even aking place on day d, he unexpeced, or surprise arge rae change can be calculaed as he change in he rae implied by he curren-monh fuures conrac, scaled up by a facor relaed o he number of days in he monh affeced by he change. In sum, we compue he unexpeced arge rae change or he surprise, as S [ D/( D d)] ( f f ) (1) d d 1 where f d is he curren-monh fuures rae a he end of he announcemen day d and D is he number of days in he monh. Kuner (2001) uses a scaled version of he one-day change in he curren-monh federal funds fuure rae because in he US he fuures conrac s payoff depends on he monhly average federal funds rae, and he scaled facor is included o reflec he number of days remaining in he monh ha are affeced by he change. This scaled facor is no required o obain a measure of he surprise in he ECB announcemen and, following Bredin e al. (2007), we proxy surprises in ECB policy raes using he one-day change in he hree-monh Euribor fuures rae. 12 The daa for he moneary policy relaed variables are provided by Bloomberg. 3. Dynamics of European governmen bond marke inegraion: Cluser Analysis Given he diversiy of economic and financial srucures across he EU economies, he fac ha no all he counries belong o he EMU and ha some counries only became members of he EU relaively recenly, i is sandard in he lieraure o divide European counries ino four groups: (1) EMU EU-15 cenral counries, (2) EMU EU-15 peripheral counries, (3) Non-EMU new EU counries, and (4) non-emu EU-15 counries. However, he recen years of urmoil migh have produced heerogeneiy wihin groups or even 12 Bernoh and Von Hagen (2004) find ha he hree-monh Euribor fuures rae is an unbiased predicor of euro area policy rae changes.
homogeneiy beween counries in differen groups in erms of heir respecive levels of inegraion. To analyse his possibiliy, we carry ou a hierarchical cluser analysis ha enables us o group counries ha presen similar characerisics across a se of variables. Here, his se of variables refers exclusively o he coexceedances of each governmen bond marke wih he euro area over ime, i.e. is level of inegraion. The end resul is a map (dendrogram) ha allows us o visualize he groups. In so doing, we are able o es wheher he cluser analysis (in erms of he level of inegraion) leads o he same classificaion of counries as described above. As menioned, from he inroducion of he common currency unil he end of 2007, he bonds of he EMU counries were almos perfec subsiues bu his siuaion changed, firs, wih he financial crisis and, subsequenly, wih he European sovereign deb crisis. For his reason, o deermine wheher he cluser groups have been sable over he sample period, we perform he cluser analysis for hree sub-samples: he ranquil period, he financial crisis and he European sovereign crisis. Figure 1 shows ha during he ranquil period, he cenral counries (Ausria, Belgium, France, Germany and he Neherlands) form a cluser, o which Ialy is added. The dendrogram also shows he similariy beween he peripheral bond markes (Greece, Ireland, Porugal and Spain), wih he excepion of Ialy, and o which Denmark, Hungary and Poland are added. Finally, he UK, Sweden and he Czech Republic form a separae alignmen. The picure changes somewha when we consider he financial crisis (Figure 2). As in he ranquil period, he cenral counries form a cluser, o which Spain and Denmark are now added. 13 The Czech Republic and Poland, he new EU members, cluser ogeher. The similariies beween Greece, Ireland and Porugal, he firs peripheral economies o collapse, can be clearly idenified and hey form a group ogeher wih Ialy. The UK and Sweden coninue o be independen of he oher counries, as now is Hungary. Finally, Figure 3 shows ha during he European sovereign deb crisis, as expeced, he cenral counries once more form a group (ogeher wih Denmark). 14 Ialy and Spain cluser ogeher, as do he new members of he EU ogeher wih Greece, Ireland and Porugal. This resul indicaes ha he peripheral bond markes are divided ino wo 13 I is no unil he second quarer of 2008 ha Spain wen ino recession (Orega and Peñalosa, 2012). 14 As Ehrmann e al. (2011) and Södersröm (2010) poin ou, Denmark s exchange rae and moneary policy are pegged so ighly o he euro and he ECB ha he counry s bonds display a very high degree of inegraion wih hose of he euro area.
groups: hose mos affeced by he European sovereign deb crisis (Greece, Ireland and Porugal), whose levels of inegraion fall o levels similar o hose of he new member saes; and Ialy and Spain, which remain a higher levels of inegraion. Sweden and he UK also form a separae alignmen. Overall, he resuls of he clusering analysis sugges ha he level of inegraion of European governmen bond markes wih he euro area, measured in erms of coexceedances, has changed over he hree sub-samples under analysis. Alhough he cenral bond markes cluser ogeher and he UK and Sweden are independen hroughou he hree sub-periods, he remaining governmen bond markes presen a cerain degree of insabiliy indicaing ha he effecs of he crises have no been homogeneous across hese counries. As a consequence, opporuniies for diversificaion have changed over he sample period. 4. Deerminans of European governmen bond marke inegraion: Logisic regression model Our aim is o idenify he underlying deerminans of he dynamics of inegraion observed and o deermine wheher heir imporance varies across counries and/or groups of counries. A coexceedance is a variable equal o one when we record an exreme reurn in a European governmen bond marke and in he euro area simulaneously on a given day and zero oherwise. As such, we can use he binomial logi model, a frequenly adoped approach for esimaing he probabiliies associaed wih evens capured in a dichoomous variable. Defining y as being equal o one when here is a coexceedance on a given day and zero oherwise, he probabiliy of a coexceedance in he binomial logi model can be given by Pr(y 1) exp(x β) /[1 exp(x β)] (2) ' ' ' where he vecor x includes he explanaory variables menioned above plus a consan and β is a vecor of coefficiens. When β i is significan, hen he variable x i affecs he probabiliy of he occurrence of a coexceedance. The model is esimaed using maximum likelihood and goodness-of-fi is measured using McFadden s (1974) pseudo-r 2 approach. We esimae he model separaely for posiive and negaive coexceedances o allow he facors o have differen effecs on each ail. As we are also ineresed in deermining
wheher he effecs of he facors differ during he financial and he European sovereign deb crises, we include an inercep dummy as well as ineracion dummies for all model variables, where he dummy variable equals one during he sovereign deb crisis (from 1 January 2010 unil he end of he sample period) and zero before. 15 Table 3 (Table 4) shows he parameer values when esimaing he binomial logi model for he negaive (posiive) coexceedance variable for each European governmen bond marke wih he euro area. To es he subsiuion effec we include boh money marke and sock marke variables. Our resuls reveal a subsiuion effec primarily beween bonds and money marke insrumens. Specifically, he likelihood of observing coexceedances is negaively relaed o he Euribor, which indicaes ha increasing ineres raes are likely o induce fligh-o-qualiy episodes, and hus, diminish inegraion. This migh be because money marke insrumens presen a lower degree of risk han bonds and his is especially imporan in periods of urmoil when invesors are ineresed in safe asses. Ineresingly, in he case of op-ail coexceedances, he subsiuion effec was recorded during he sovereign deb crisis, while in he case of boom-ail coexceedances, he subsiuion effec ook place during he financial crisis. As for he sock marke, evidence in favour of a subsiuion effec beween sock and bond markes is scarce and heerogeneous. In general, our resuls confirm he hypohesis ha inegraion increases in highly volaile periods in he regional marke. The likelihood of posiive coexceedances increases during boh crises while, in he case of negaive coexceedances, he likelihood increases only during he sovereign deb crisis. As for differences across counries, he cenral bond markes are hose ha show mos evidence in favour of high-frequency propagaion of shocks in a volaile European environmen. In conras, he resuls from he new members bond markes fail o suppor his hypohesis. An examinaion of he impac of unexpeced news announcemens released by he ECB shows ha hey only appear o be useful in explaining negaive coexceedances in he case of cenral and non EMU EU-15 bond markes during he sovereign deb crisis, and in he case of Greece, Ialy and he UK during he financial crisis. In line wih Abad and Chuliá (2013), his resul suggess ha unexpeced news releases from he ECB increase uncerainy and decrease he level of inegraion of hese bond markes wih he euro area. 15 As he number of coexceedances during he ranquil period is almos zero, he binomial logi regression analysis is carried ou only during he crisis periods.
In addiion, we examine wheher some fracion of he coexceedances of each bond marke wih he euro area can be explained by he explanaory variables associaed wih he US, i.e. wheher global facors have an effec on he inegraion of European governmen bond markes. 16 Our resuls do no show a subsiuion effec beween he US and European asses considered; however, as he US sock marke becomes more volaile, he more likely we are o observe a rise in he level of inegraion of European governmen bond markes. As in he case of regional volailiy, here is no evidence of his effec among he new members. Ineresingly, in he case of op-ail coexceedances, he volailiy effec is recorded during he financial crisis while in he case of boom-ail coexceedances, i is recorded during he sovereign deb crisis bu in fewer bond markes. Finally, he regression coefficiens for he moneary policy surprises announced by he Fed are insignifican during boh he financial and he sovereign deb crises and for boh boom- and op-ail coexceedances. Finally, our analysis of he facors associaed wih he inegraion of European governmen bond markes only reveals differences beween new members and he EU-15 members in erms of he impac of unexpeced news announcemens from he ECB and of boh European and US volailiy. 5. Conclusions Using he coexceedance measure proposed by Bae e al. (2003), we have analysed he degree of inegraion of European governmen bond markes wih he euro area. This approach has allowed us o adop an inuiive measure of inegraion: he higher he number of coexceedances of each bond marke wih he euro area, he higher he degree of inegraion. In a firs sep, we carried ou a hierarchical cluser analysis ha allowed us o analyse he way in which he bond markes group over he sample period (comprising a ranquil period, he financial crisis and he sovereign deb crisis) in erms of he degree of inegraion. In a second sep, we have used a binomial logisic regression model o deermine he facors associaed wih an increase (decrease) in he probabiliy of observing a coexceedance. Specifically, we were ineresed in esing wheher (i) here is a subsiuion effec beween equiies, money marke insrumens and bonds, (ii) he propagaion of shocks is more likely in a highly volaile environmen, (iii) moneary policy surprises 16 Owing o iming convenions (European markes close before heir US counerpar), US explanaory variables ener he model lagged one period. We inerpre hese resuls as evidence of he predicabiliy of coexceedances.
announced by he ECB decrease he level of inegraion of European governmen bond markes, and (iv) European inegraion is also driven by inernaional facors. We repor evidence ha he degree of inegraion of European governmen bond markes has changed over he sample period: firs, during he financial crisis and, subsequenly, during he European sovereign deb crisis. Moreover, he effecs of hese crises have no been he same across all he bond markes, o he exen ha he radiional groupings of markes on he basis of he level of inegraion vary across he hree sub-samples. In he case of he facors associaed wih he likelihood of he occurrence of coexceedances, we obain a number of ineresing resuls. For example, our findings poin o a subsiuion effec beween bonds and European money marke insrumens, bu no beween bond and sock markes (in eiher he US or Europe). As expeced, in urbulen imes he subsiuion effec involves he leas risky asse. In addiion, we find evidence indicaing ha he greaer he volailiy in European and US financial markes, he more likely we are o observe he propagaion of shocks in boh ails. Finally, our resuls show ha unexpeced news announcemens from he ECB increase uncerainy and weaken he degree of inegraion of European governmen bond markes. Our resuls should enable marke paricipans o make effecive invesmen decisions, given ha hey need o have an undersanding of he way in which exreme shocks propagae across European governmen bond markes. Addiionally, our findings should be of use o policymakers as hey srive o undersand he effecs of heir moneary policy decisions on bond markes in imes of exreme shocks. Acknowledgemens This work has been funded by he Spanish Minisry of Economy and Compeiiveness (ECO2011-23959 and ECO2012-35584).
6. References Abad, P. and H. Chuliá. European Governmen Bond Markes and Moneary Policy Surprises: Reurns, Volailiy and Inegraion, Working paper 25, Research Insiue of Applied Economics (IREA), 2013. Abad, P., Chuliá, H. and M. Gómez-Puig. Time-varying inegraion in European Governmen Bond Markes. European Financial Managemen, 2014, 20(2), 270-290. Abad, P., Chuliá, H., and M. Gómez-Puig. EMU and European governmen bond marke inegraion. Journal of Banking and Finance, 2010, 34, 2851 2860. Bae, K.-H., Karolyi, G.A. and R.M. Sulz. A new approach o measuring financial conagion. Review of Financial Sudies, 2003, 16 (3), 717 763. Baur, D. and N. Schulze. Coexceedances in Financial Markes - A Quanile Regression Analysis of Conagion. Emerging Markes Review, 2005, 6(1), 21-43. Beber, A., Brand, M. W. and K.A. Kavajecz. Fligh-o-Qualiy or Fligh-o-Liquidiy? Evidence from he Euro-Area Bond Marke. Review of Financial Sudies, 2009, 22(3), 925.957. Bernanke, B.S. and Kuner, K.N. Wha explains he sock marke s reacion o Federal Reserve policy? Journal of Finance, 2005, 60 (3), 1221 1257. Bernoh, K. and J. Von Hagen. Euribor fuures marke: efficiency and he impac of ECB policy announcemens. Inernaional Finance, 2004, 7, 1-24. Bomfim, A. N. Pre-announcemen effecs, news effecs, and volailiy: Moneary policy and he sock marke. Journal of Banking & Finance, 2003, 27(1), 133-151. Bredin, D., Hyde, S., Nizsche, D. and G. O Reilly. European Moneary Policy Surprises: The aggregae and secoral sock marke response. Inernaional Journal of Finance and Economics, 2007, 14(2), 156-171. Brunnermeier, M. K., and L. H. Pedersen. Marke liquidiy and funding liquidiy. Review of Financial sudies, 2009, 22(6), 2201-2238. Caballero, R.J. and A. Krishnamurhy. Collecive risk managemen in a fligh o qualiy episode. Journal of Finance, 2008, 63, 2195 2230. Cappiello, L., Gerard, B., Kadarenja, A. and S. Mangenelli, S. Financial Inegraion of New EU Member Saes, Working Paper, ECB, 2006. Cassola M. and Morana, C. Euro money marke spreads during he 2007 financial crisis. Journal of Empirical Finance, 2012, 19(4), 548-557.
Chrisiansen, C., Volailiy-spillover effecs in European bond markes. European Financial Managemen, 2007, 13(5), 923-948. Chrisiansen, C. and Ranaldo, A. Exreme Coexceedances in New EU Member Saes Sock Markes. Journal of Banking and Finance, 2009, 33(6), 1048-1057. Chrisiansen, C. Inegraion of European Bond Markes. Journal of Banking and Finance, 2014, 42, 191-198. Dungey, M. and Marin, V. Unravelling Financial Marke Linkages During Crises. Journal of Applied Economerics, 2007, 22(1), 89-119. Ehrmann, M., Frazscher, M., Gürkaynak, R. S. and Swanson, E. T. Convergence and anchoring of yield curves in he euro area. The Review of Economics and Saisics, 2011, 93(1), 350-364. Geyer, A., Kossmeier, S. and S. Pichler. Measuring sysemaic risk in EMU governmen yield spreads. Review of Finance, 2004, 8, 171 197. Gómez-Puig, M. The immediae effec of moneary union over EU-15 s sovereign deb yield spreads. Applied Economics, 2009a, 41, 929 939. Gómez-Puig, M. Sysemic and idiosyncraic risk in EU-15 sovereign yield spreads afer seven years of Moneary Union. European Financial Managemen, 2009b, 15, 971 1000. Kim, S-J., Lucey, B. and E. Wu. Dynamics of bond marke inegraion beween esablished and accession European Union counries. Journal of Inernaional Financial Markes, Insiuions and Money, 2006, 16(1), 41-56. Kuner, K. Moneary policy surprises and ineres raes: evidence from he fed funds fuures marke. Journal of Moneary Economics, 2001, 47(3), 523-544. Manganelli, S. and Wolswijk, G. Wha drives spreads in he euro area governmen bond marke? Economic Policy, 2009, 24(4), 191 240. McFadden, P. The Measuremen of Urban Travel Demand. Journal of Public Economics, 1974, 3, 303--328. Orega, E. and Peñalosa, J. The Spanish economic crisis: key facors and growh challenges in he euro area. BDE occasional paper nº. 1201, BDE, 2012. Pagano, M., and E.L. von Thadden. The European bond markes under EMU. Oxford Review of Economic Policy, 2004, 20, 531 554. Pozzi, L. and Wolswijk, G. The Time-Varying Inegraion of Euro Area Governmen Bond Markes. European Economic Review, 2012, 56(1), 36-53.
Södersröm, U. Re-Evaluaing Swedish Membership in he EMU: Evidence from an Esimaed Model. In Albero Alesina and Francesco Giavazzi (Eds.), Europe and he Euro. Chicago: Universiy of Chicago Press, 2010. Von Hagen, J., Schuknech, L. and Wolswijk, G. Governmen bond risk premiums in he EU revisied: The impac of he financial crisis. European Journal of Poliical Economy, 2011, 27(1), 36-43.
6. Tables Table 1. Coexceedances: Relaive Frequency Coexceedances Negaive coexceedances Posiive coexceedances Enire Sample Sample Sample Enire Sample Sample Sample Enire Sample Sample Sample sample A B C sample A B C sample A B C Panel a) EMU EU-15 Cenral Ausria 0.068 0.025 0.116 0.066 0.035 0.013 0.059 0.034 0.033 0.012 0.057 0.031 Belgium 0.074 0.028 0.140 0.064 0.037 0.015 0.065 0.035 0.037 0.013 0.075 0.029 France 0.068 0.027 0.115 0.067 0.035 0.015 0.056 0.036 0.033 0.012 0.059 0.030 Germany 0.060 0.024 0.108 0.054 0.030 0.015 0.048 0.028 0.029 0.009 0.061 0.022 Neherlands 0.066 0.030 0.118 0.057 0.034 0.015 0.057 0.032 0.031 0.015 0.061 0.023 Panel b) EMU EU-15 Peripheral Greece 0.027 0.004 0.062 0.020 0.009 0.000 0.024 0.006 0.014 0.004 0.038 0.006 Ireland 0.029 0.006 0.064 0.023 0.013 0.004 0.026 0.011 0.014 0.001 0.037 0.008 Ialy 0.050 0.027 0.081 0.046 0.023 0.015 0.040 0.018 0.026 0.012 0.041 0.025 Porugal 0.025 0.001 0.045 0.029 0.012 0.001 0.024 0.012 0.011 0.000 0.021 0.013 Spain 0.044 0.006 0.086 0.043 0.020 0.006 0.040 0.018 0.023 0.000 0.046 0.024 Panel c) non-emu new EU Czech Republic 0.020 0.007 0.035 0.019 0.005 0.003 0.006 0.005 0.007 0.004 0.006 0.010 Hungary 0.024 0.001 0.051 0.023 0.006 0.001 0.008 0.007 0.005 0.000 0.003 0.009 Poland 0.021 0.004 0.035 0.022 0.006 0.004 0.005 0.007 0.005 0.000 0.005 0.008 Panel d) non-emu EU-15 Denmark 0.044 0.001 0.070 0.055 0.023 0.001 0.035 0.029 0.020 0.000 0.035 0.023 Sweden 0.028 0.007 0.040 0.035 0.013 0.004 0.019 0.016 0.010 0.003 0.011 0.014 UK 0.033 0.007 0.067 0.029 0.016 0.001 0.037 0.013 0.014 0.006 0.026 0.012 Noe: Sample A refers o he ranquil period exending form 1 January 2005 o 6 Augus 2007. Sample B refers o he financial crisis period exending from 7 Augus 2007 o 31 December 2009. Sample C refers o he European sovereign deb crisis exending from 1 January 2010 o 15 December 2013.
Table 2. ANOVA es of mean equaliy Coexceedances Negaive coexceedances Posiive coexceedances B C samples A B C samples B C samples A B C samples B C samples A B C samples Panel a) EMU EU-15 Cenral Ausria 12.728* 22.790* 5.854* 10.296* 6.859* 11.700* (0.000) (0.000) (0.016) (0.000) (0.009) (0.000) Belgium 27.187* 33.185* 8.137* 11.982* 18.595* 20.307* (0.000) (0.000) (0.004) (0.000) (0.000) (0.000) France 11.454* 21.164* 3.689* 8.223* 8.270* 12.608* (0.001) (0.000) (0.055) (0.000) (0.004) (0.000) Germany 17.267* 22.585* 4.380* 6.275* 16.185* 18.057* (0.000) (0.000) (0.037) (0.002) (0.000) (0.000) Neherlands 19.594* 23.019* 6.270* 9.252* 15.124* 14.069* (0.000) (0.000) (0.012) (0.000) (0.000) (0.000) Panel b) EMU EU-15 Peripheral Greece 19.670* 23.569* 10.226* 11.974* 23.335* 18.783* (0.000) (0.000) (0.001) (0.000) (0.000) (0.000) Ireland 17.311* 20.549* 5.344* 5.585* 17.876* 18.151* (0.000) (0.000) (0.021) (0.004) (0.000) (0.000) Ialy 8.917* 11.111* 6.912* 5.366* 3.356* 6.282* (0.003) (0.000) (0.009) (0.005) (0.067) (0.002) Porugal 2.763* 13.775* 3.649* 7.332* 1.649 6.917* (0.097) (0.000) (0.056) (0.001) (0.199) (0.001) Spain 13.222* 25.812* 7.729* 9.561* 5.956* 16.419* (0.000) (0.000) (0.006) (0.000) (0.015) (0.000) Panel c) non-emu new EU Czech Republic 3.857* 6.866* 0.166 0.480 0.519 0.953 (0.05) (0.001) (0.684) (0.619) (0.471) (0.386) Hungary 9.182* 17.873* 0.074 1.628 1.825 3.713* (0.003) (0.000) (0.786) (0.197) (0.177) (0.025) Poland 2.384 8.338* 0.265 0.306 0.529 2.781* (0.123) (0.000) (0.607) (0.736) (0.467) (0.062) Panel d) non-emu EU-15 Denmark 1.478 21.863* 0.447 9.613* 1.990 11.700* (0.224) (0.000) (0.504) (0.000) (0.159) (0.000) Sweden 0.259 7.631* 0.300 2.618* 0.186 2.687* (0.611) (0.001) (0.584) (0.073) (0.666) (0.068) UK 13.478* 18.776* 10.637* 12.995* 4.499* 5.254* (0.000) (0.000) (0.001) (0.000) (0.034) (0.005) Noe: F-es denoes he ANOVA es wih he null hypohesis ha he number of coexceedances is equal in he A, B and C (B and C) sub-samples (see noe o Table 1). * indicaes ha he null hypohesis is rejeced.
Table 3. Parameer esimaes from he binomial logi model for he posiive coexceedance variable Consan ECB FED EA US EA US C SEA, SUS, SR SR R R 1 S S 1 S 1 Ausria Belgium -3.484-3.11-0.076 0.062-0.003 0.103* 7.522 3.869 2.120 0.125-46.265* -5.235-12.192-12.732* 21.06* -13.159 810.78* 554.851 1765.034* 1874.804* France -3.433-0.085-0.013 4.000 0.243-17.347-7.597-5.995 847.892* 1542.855* Germany -3.545-0.088-0.039 7.579-0.37-22.149-9.085 3.493 1014.349* 2298.313* Neherlands -3.507-0.088-0.043 3.107-0.208-1.889-7.754-18.832 947.572* 1808.897* Greece Ireland -3.778-3.675-0.02-0.097 0.031 0.043-7.911-4.403 2.871* 2.407-23.204* -34.192* -5.855-8.491 13.049-4.936 520.169-15.342 1356.211 342.036 Ialy -3.83 0.129 0.017 6.07-0.791-27.688-4.484 17.249 873.121* 2030.891* Porugal -3.976-0.119 0.004 12.618-0.622-29.866-8.480 13.114 1119.652* 2777.628* Spain -5.137-0.199 0.041 13.056 1.298 5.023 7.313-43.617* 1158.995* 2368.994* Czech Hungary -6.502-6.652 0.036-0.045-0.300-0.294-22.375-8.328-10.872* 0.249 7.741 13.016 43.898* -26.528 0.342-18.746 71.679-336.184 360.901 4300.496 Poland -7.268 0.077-0.527-1.153-0.659-7.056 32.969 30.414 1793.26* 880.664 Denmark Sweden -4.029-5.588-0.139-0.032 0.025-0.320 1.668 1.855 1.725 2.619 3.145-32.354-2.727-17.263-36.906* 34.43 364.11 900.071 2270.556* 2993.231* UK -4.368 0.116 0.075 7.212 1.170-47.943-17.913* 22.291 393.296 2216.545* ECB FED EA US EA US C SEA, SUS, D D SR D SR D R D R 1 D S D S 1 D S D D 1 MF R 2 Ausria Belgium -1.022* -1.373* 0.157-0.128 0.567 0.511-38.793* -44.952* 4.363 19.99-4.575-30.219 3.745 8.704 2.523 45.721 3248.724* 3119.28* 2520.22 2152.571 0.143 0.134 France -1.005* 0.155-0.108-39.111* 13.346 56.261 11.414-63.701 2535.65* 1327.275 0.123 Germany -1.365* 0.077-0.124-33.049* 13.503 23.8-0.247-61.99 2065.85 218.11 0.182 Neherlands -1.416* 0.081-0.012-33.237* 14.904-40.634-5.41 6.098 2410.61* 1445.982 0.178 Greece Ireland -2.663* -2.451* 0.025 0.333 0.322 0.034-44.871-45.991 66.485* 41.528 51.865* 105.287* 0.893 15.467 14.446-24.338 215.806 2250.288-4003.923 294.287 0.172 0.174 Ialy -1.174* -0.224 0.742-41.147* 18.486-3.419 150.926* 1874.827 631.129 0.158 Porugal -0.525-0.016 0.732-45.137* 13.621-7.773 25.836 47.78 1185.944-1433.581 0.199 Spain -0.48 0.176 1.048-41.294 14.674-42.704 8.6 106.417* 2943.911* 883.659 0.140 Czech Hungary 0.196 0.815 0.203 0.091 0.254-0.377-46.808-79.073* 80.486* 33.857 16.216 84.768* -52.293 55.554 93.514* 71.306 1252.511-5450.643 1848.029-9956.382 0.289 0.339 Poland 1.395-0.053 0.175-61.121 9.72 96.456* 1.038 7.412-4509.754-4288.565 0.311 Denmark Sweden -0.939* -0.196 0.144 0.012-0.227-0.26-38.389* -55.88* 10.679 17.197-29.731 47.745-3.034 11.483-15.412-93.561* 2646.257* 3756.226* 400.96 2389.205 0.195 0.203 UK -2.056* -0.182-0.656 5.802 29.114 57.882 12.22-4957.798* 3541.887 0.235 ECB FED EA US Noe: * indicaes significance a he 10% level. SR and SR 1 refer o moneary policy surprises announced by he ECB and he Fed, respecively; R and R 1 refer o he 3- EA US C monh inerbank ineres rae (Euribor) and he 3-monh Treasury bill rae, respecively; S, S 1 and S refer o he Eurosoxx50 index reurns, he S&P500 index reurns and, he sock index reurns of each counry, respecively; and S EA SUS, and 1 refer o he volailiy of he Eurosoxx50 index reurns and he S&P500 index reurns, respecively. Volailiy series have been muliplied by 100. D refers o a dummy variable ha equals one during he sovereign deb crisis (from 1 January 2010 unil he end of he sample period) and zero before. MF R 2 refers o McFadden s pseudo-r 2.
Table 4. Parameer esimaes from he binomial logi model for he negaive coexceedance variable Consan ECB FED EA US EA US C SEA, SUS, SR SR R R 1 S S 1 S 1 Ausria Belgium -3.084-3.085-0.124-0.111-0.394-0.423-6.596-11.658-1.985-1.897 34.327* -2.126 1.182 3.429-0.168 47.059* 148.338 45.759-157.705 720.765 France -3.182-0.053-0.393-8.643-2.715 35.065 1.652 2.574 99.702-184.228 Germany -3.393-0.062-0.422-14.215-2.185 30.034 4.589 0.412 123.937 475.328 Neherlands -3.121-0.052-0.371-11.025-2.576 19.237-1.911 16.425 2.538-135.368 Greece Ireland -4.521-4.672-0.209* -0.11-0.425-0.481-31.563* -25.906* -3.672 2.491 16.319 9.673 17.532 3.969-9.038 42.625* 331.598 269.642-194.819-465.153 Ialy -3.79-0.159* -0.365-15.835-3.304* -9.351 5.595 30.333 463.947 520.441 Porugal -3.628-0.087-0.423-16.191-1.95 8.553 3.744 18.325 225.412 283.695 Spain -4.083-0.113-0.556-31.479* 2.09 8.161 4.173 28.197-389.348-390.693 Czech Hungary -5.914-6.136 0.011-0.155-0.329-0.253-35.757* -22.414-6.202* -5.03* 6.496-24.365 3.320 16.805-7.652-19.182-335.709 239.952 1410.031-42.665 Poland -8.858-0.061-0.334-82.38* -0.06-3.592 15.711 - -1560.019-3268.738 Denmark Sweden -3.66-5.09-0.094-0.14-0.453-0.512-22.89* -12.025 0.855 1.532 37.108* 31.498-1.864-13.326-20.872-14.132-167.627 761.85 586.471 2602.408* UK -3.844-0.259* -0.394-23.485* -3.904* 20.985 4.693-5.863 203.992 73.573 ECB FED EA US EA US C S, EA SUS, D D SR D SR D R D R 1 D S D S 1 D S D D 1 MF R 2 Ausria Belgium -0.894* -0.939* -0.208-0.186 0.447 0.438 17.2 5.927 10.15 15.645-22.483 30.96-15.011-1.146 13.479-71.938* 2541.701* 2891.23* 4336.641* 2183.815 0.096 0.103 France -0.904* -0.254* 0.448 1.7 13.907-120.887-0.317 104.015 2868.618* 2705.561 0.100 Germany -0.938* -0.309* 0.418 22.756 23.149-53.961-6.468 72.571 2376.541* 5020.625* 0.123 Neherlands -0.917* -0.283* 0.314 17.231 15.966-32.644 5.333 38.815 2348.179* 4440.557* 0.106 Greece Ireland -1.862* -0.018 0.216 0.028 0.504 0.261 58.405 70.504* -40.569-17.668-30.744-31.127-23.771 29.411 17.472-38.898 4534.3* 351.286 3213.252 2515.842 0.183 0.163 Ialy -0.802-0.116 0.32 12.557-0.376 8.334 4.032-67.069 977.221-71.779 0.128 Porugal -0.774-0.230 0.395 18.596 4.380-34.024-6.053-32.288 444.564 416.436 0.107 Spain -0.968 0.024 0.270 83.201* -18.564 14.655-15.778-81.324* 2975.186* 4476.486 0.105 Czech Hungary -0.096 0.21-0.14-0.252-0.292 0.276 94.692* 20.755-20.25-16.063-88.247* -9.85 30.086-43.29-2.505-13.388-811.582 1230.489-3528.975 1150.414 0.176 0.230 Poland 3.244-0.557 0.041 126.500-52.125-18.072 22.210-19.104-4218.89-2667.215 0.432 Denmark Sweden -0.61 0.784-0.28* -0.298* 0.456 0.491 27.28 28.308 10.524-5.909-25.574-40.936 12.619 30.492 56.543* 53.05 2758.365* -803.114 3213.08-467.842 0.104 0.150 UK -1.451* 0.189 0.254 62.404* 45.785* -21.286-16.234 79.834 2029.392 5809.845* 0.158 Noe: See noe o Table 3.